We will see...
I believe that you have not understood each other very well this post having created with a focus different to what you usually interpret...
The experiment has been useful and amusing.
1 - this evaluation is very useful to compare the evolution of different versions of the same engine, to compare among ryka 1, rybka 2, rybka 3 stiles, rybka mp. and to compare among fritz 6 to fritz 11, and versions deep... and to compare chesstiger versions... and to compare hiarsc 9 to 12...
2 - when we block analysis to 1 ply, at least we make sure of understanding the first movements candidates of the program, without influencing the quick cpus and code optimizations, perhaps be not exactly 1 ply, but it will always be the same analysis depth, and that is the important thing..
3 - this will reveal concept shortcomings in some positions...
4 - it can improve the level of an engine improving the selections that he/she makes in the first play... (first movements candidates) because starting from here the whole tree of variants is born.
5 - other engines calculate more +1 ply too...= conditions...
I am carrying out an experiment among knowledge in chess and search depth.
(it seems to be that everybody this in agreement in that rybka calculates plys info + 2 or ply +3, I am not so sure of that, I think that this perhaps is certain when they are analyzing 15 or 20 plys, but I don't believe that it is exactly that in the level = 1 ply, anyway that is the less important thing, when we are evaluating the same maker's engine, or in case other engines also carries out an approach to the rise. I think that other engines also makes those things... ; -))
Also, that information about rybka, streka (...base fruit, etc...) he/she doesn't explain as rybka 3 human (the best engine in the universe,; -)) it loses 2-0 ostentatiously against an engine of ago 3 or 5 years. (chesstiger 14 are the engine more veteran of those that I even use... I like their style. )
see the two games here: (I try that plays a match of 12 departures, but the departures were duplicated an and another time, that which has much sense, when analysis is used for it limits of plys, and these they are so low that they don't allow different deviations of the main branches. )
Here the victory game for Chesstiger 15:
Well, looking at this departure where rybka 3 human loses, we cannot see as rybka this thinking 1 single ply, such and as we understand it the humans, but that is not outstanding for that chesstiger neither him this making...[Event "PC-820D-OLIV, 1Ply / 1Ply"]
[Site "PRIVADA"]
[Date "2008.09.05"]
[Round "1"]
[White "Rybka 3 Human 1-cpu 32-bit kz"]
[Black "Chess Tiger 15.0"]
[Result "0-1"]
[ECO "A06"]
[WhiteElo "2600"]
[BlackElo "2600"]
[PlyCount "100"]
[EventDate "2008.??.??"]
[TimeControl "120"]
1. Nf3 {Ambos ultima jugada del libro 0.09/1 0} d5 {-0.36/3 1} 2. d3 {
(Nc3) 0.18/1 0} e6 {(Nf6) -0.28/3 0} 3. Nc3 {0.28/1 0} Nf6 {(Nc6) -0.20/3 0} 4.
Be3 {(e3) 0.24/1 0} Nc6 {-0.42/3 0} 5. Nb5 {(d4) 0.09/1 0} a6 {-0.68/3 0} 6.
Nbd4 {0.09/1 0} Nxd4 {(Ng4) -0.56/3 0} 7. Bxd4 {-0.10/1 0} c5 {(Bd6) -0.38/3 0}
8. Be5 {-0.14/1 0} Bd6 {-0.38/3 0} 9. Qd2 {(e3) -0.20/1 0} h6 {(0-0) -0.38/3 0}
10. Qf4 {(e3) -0.02/1 0} Bxe5 {-0.40/3 0} 11. Nxe5 {-0.06/1 0} Qe7 {
(0-0) -0.26/3 0} 12. Qg3 {(e4) -0.03/1 0} O-O {-0.32/3 0} 13. e3 {-0.11/1 0} b6
{(Nd7) -0.20/3 0} 14. Be2 {0.18/1 0} Qe8 {(Nd7) -0.08/3 0} 15. O-O {
(a3) 0.61/1 0} Ra7 {(Nd7) -0.08/3 0} 16. f4 {0.71/1 0} Qb5 {(Bb7) -0.14/3 0}
17. Rfb1 {(b3) 0.57/1 0} Qa4 {(Ne4) -0.18/3 0} 18. b3 {0.59/1 0} Qe8 {
(Qa3) -0.06/3 0} 19. b4 {(a4) 0.78/1 0} cxb4 {(Qe7) -0.34/3 0} 20. Rxb4 {
0.96/1 0} b5 {(Rc7) 0.18/3 0} 21. Rc1 {(a4) 0.74/1 0} a5 {(Rc7) -0.02/3 0} 22.
Rbb1 {(Rb3) 0.76/1 0} Ba6 {(Rc7) 0.04/3 0} 23. c3 {0.80/1 0} Rc7 {0.18/3 0} 24.
a4 {(Bg4) 0.60/1 0} bxa4 {-0.36/3 0} 25. Rb6 {0.34/1 0} Ne4 {(Qc8) -0.68/3 0}
26. dxe4 {0.66/1 0} Bxe2 {-0.16/3 0} 27. exd5 {-0.14/1 0} exd5 {-0.54/3 0} 28.
Rc2 {(Kf2) -0.14/1 0} Bh5 {(f6) -0.64/3 0} 29. f5 {(Rxh6) -0.17/1 0} f6 {
-0.68/3 0} 30. Nd3 {(Ng4) -0.41/1 0} Re7 {(a3) -1.28/3 0} 31. Qd6 {-0.66/1 0}
Rxe3 {(Bf7) -1.72/3 0} 32. Qxd5+ {-1.26/1 0} Bf7 {-1.48/3 0} 33. Qd4 {-1.26/1 0
} Re4 {(Qe4) -1.40/3 0} 34. Qc5 {-1.03/1 0} Bc4 {(Re1+) -1.44/3 0} 35. Rd2 {
-0.74/1 0} a3 {(Qe7) -2.00/3 0} 36. Rd6 {(g3) -1.52/1 0} a2 {-2.52/3 0} 37. Rd1
{-2.21/1 0} Qa4 {(Qb5) -2.84/3 0} 38. Ra1 {-2.48/1 0} Rb8 {-3.40/3 0} 39. Nc1 {
(Nf2) -3.39/1 0} Re1+ {-4.60/3 0} 40. Kf2 {-3.39/1 0} Re5 {-4.78/3 0} 41. Qd4 {
(Qa7) -3.57/1 0} Rxf5+ {-5.44/3 0} 42. Kg3 {-3.52/1 0} Rb2 {-4.60/3 0} 43. h4 {
(Rd8+) -3.48/1 0} Rf1 {(Qb5) -5.58/3 0} 44. Kh2 {-3.30/1 0} Re2 {
(Re1) -4.10/3 0} 45. Nxe2 {(Qg4) -0.47/1 0} Rxa1 {-2.58/3 0} 46. Rd8+ {
-1.68/1 0} Kf7 {-2.58/3 0} 47. Qa7+ {-2.11/1 0} Kg6 {-5.50/3 0} 48. h5+ {
(Nf4+) -3.18/1 0} Kh7 {-6.50/3 0} 49. Kg3 {(Qb8) -6.16/1 0} Bxe2 {-10.06/3 0}
50. c4 {(Rd4) -7.46/1 0} Re1 {(Bxh5) -15.36/3 0} 0-1
[d]4qrk1/2r2pp1/b3pn1p/pp1pN3/5P2/2PPP1Q1/P3B1PP/1RR3K1 w - - 0 24
Now, rybka will play a4, what doesn't stop to be a surprising play for 1 ply. ¡ (lies and more lies,; -))
Now then, we see that few plays later chesstiger neither gives the impression of being very Christian in its evaluations,:
[d]4qrk1/2r2pp1/bR2pn1p/p2pN3/p4P2/2PPP1Q1/4B1PP/2R3K1 b - - 0 25
Ce4 (Ne4) is this position to be surprising, if alone we think of 1 ply of depth.
At the end everything decreases to see that engine lies better... ; -)
http://img384.imageshack.us/my.php?imag ... iesgd0.jpg
------------
After this whole amusing party of calculations of little depth, rybka loses, because its dynamic sacrifices, they are not supported with the enough power of calculate, because we have limited its capacity to 1 ply (ok, perhaps be not 1, but that that if we know it is that they are very few, the enough ones to "understand" that this valuing the algorithm of chess knowledge)
this way the influence of the optimization of the code and the speed of the cpu doesn't hide the valuations of the main algorithm.
to say it in few words, that first evaluation is the most important, because of her they leave all the other ones. It is a recursive function.
Therefore, I deduce that rybka needs of certain cpu power to have good results, because "its first analysis" is detailed.
(this has much sense if we check like in a computer the phoenix resurrection and revelation computers, where the cpu is a processor ARM to 500 Mhz speed, rybka doesn't obtain so good results as we wait, in fact, I have read for the forum that rybka needs of enough time to get good results, Deep Sjeng 3.0 get a better result when the cpu is slower, winning tournament: ... see here:)
http://www.phoenixcs.nl/index.php?optio ... 3&Itemid=1
Conclusion, when limited to 1 ply the capacity of depth is transforming our cpu into a very very slow cpu, then we can deduce that engine has more chess knowledge, (strategic valuations and position, and with little influence tactics)
This you confirms when facing all the rybka modules among them, rybka 1, rybka 2, rybka 3, rybka 3 styles.... Wisely we see as the result of the tournament it is exactly the prospective one, the most modern engines is orderly perfectly, what means that rybka goes acquiring more knowledge of version chess in version.
In the same way, the engine that he/she obtains better results is the human version what also makes sense, when limiting the search capacity triumphs to the engine that thinks of more human concepts, position them and strategic.
It is deduced then that in slow processors, the versions r-humans will be more effective.
(In other words, evaluating all the versions of the same program gets coherent results)
But I go farther...
Although rybka lies us in the real depth of analysis, it is not less certain than other engines they also lie us, then we are in equality of conditions. (~aprox)
Also although rybka really has a slight advantage, I eat it is only half advantage play in many positions an engine it can get better results if he/she has more exact knowledge of a certain position. That is observed clearly in some positions, where the rybka valuation is very incorrect.
The conclusion that I obtain is that many things can be deduced using the lowest levels in analysis, one can deduce much of the main algorithm of analysis. (chess knowledge)
Perhaps one of the rybka secrets is that it "selects-candidate-moves" their first plays with much care (carefully), because it is starting from these positions of where the main tree is born.
This has a cost important cpu clocks and for that reason rybka needs of modern computers to reflect its advantage, in slow cpu that cost in the first "selects-candidate-moves" is very expensive and if he/she doesn't have the enough capacity of depth rybka he/she makes chancy decisions that then cannot support tactically, when being much more fence the horizon of final.
Many programmers should revise like they select the first "selects-candidate-moves" and test it with problemas and diverse positions, and compary with rybka their engines, it is very important to leave of good ideas initials, although these they imply a bigger cpu cost.
-----------
For example a button: (it is a made sentence... a proverb)
Here: Rybka style dinamic slaughters (it razes) completely to Rybka style cpu in only 17 moves ¡ and r-c resings .
reason?
[Event "PC-820D-OLIV, 1Ply / 1Ply"]
[Site "PRIVADA"]
[Date "2008.09.05"]
[Round "1"]
[White "Rybka 3 Dynamic 1-cpu 32-bit"]
[Black "Rybka 3 1-cpu 32-bit"]
[Result "1-0"]
[ECO "A00"]
[WhiteElo "2600"]
[BlackElo "2600"]
[PlyCount "33"]
[EventDate "2008.??.??"]
[TimeControl "120"]
1. Nf3 {Ambos ultima jugada del libro 0.09/1 0} Nf6 {0.26/1 0} 2. Nc3 {0.09/1 0
} Nc6 {0.13/1 0} 3. d3 {0.09/1 0} d6 {0.06/1 0} 4. h3 {0.09/1 0} a6 {
(Be6) 0.07/1 0} 5. Be3 {0.12/1 0} h6 {(Be6) 0.07/1 0} 6. g4 {(a3) 0.09/1 0} Be6
{0.02/1 0} 7. Bg2 {0.13/1 0} h5 {0.10/1 0} 8. g5 {0.21/1 0} Nd7 {(Nd5) 0.20/1 0
} 9. Qd2 {(Nd2) 0.28/1 0} Nc5 {(b5) 0.27/1 0} 10. d4 {(0-0-0) 1.90/1 0} b6 {
2.04/1 0} 11. dxc5 {1.96/1 0} bxc5 {2.07/1 0} 12. g6 {(Nd5) 2.10/1 0} fxg6 {
2.53/1 0} 13. Ng5 {2.62/1 0} Bd7 {2.60/1 0} 14. Be4 {(Bd5) 3.55/1 0} Rh6 {
3.70/1 0} 15. Qd5 {4.97/1 0} e6 {5.53/1 0} 16. Nxe6 {5.04/1 0} Nb4 {5.74/1 0}
17. Qxa8 {(Nxd8) 5.99/1 0} 1-0
for that like we have limited to 1 ply (or <3 ply) the search depth, wins the engine that uses more chess knowledge...
that an engine loses in 17 plays it is not any surprise, if we don't allow it to think enough, the surprising thing is that an engine wins in 17 plays...
----------
To teach as other engines they also have a valuation different to 1 ply, but anyway a low evaluation sees the two games that rybka could not win to an engine that is possibly 5 years old, I don't remember it exactly, but although ch14 is old, the chess knowledge don't expire...
[Event "PC-820D-OLIV, 1Ply / 1Ply"]
[Site "PRIVADA"]
[Date "2008.09.05"]
[Round "2"]
[White "Chess Tiger 14.0"]
[Black "Rybka 3 Human 1-cpu 32-bit kz"]
[Result "1/2-1/2"]
[ECO "A05"]
[WhiteElo "2600"]
[BlackElo "2600"]
[PlyCount "118"]
[EventDate "2008.??.??"]
[TimeControl "120"]
1. Nf3 {Ambos ultima jugada del libro 0.09/1 0} Nf6 {0.23/1 0} 2. e3 {
(Nc3) 0.48/3 0} Nc6 {0.05/1 0} 3. d4 {(Nc3) 0.52/3 0} d6 {(d5) -0.09/1 0} 4.
Bc4 {(d5) 0.48/3 0} Bf5 {(e6) 0.07/1 0} 5. Bd3 {(0-0) 0.44/3 0} Qd7 {
(e6) -0.09/1 0} 6. Bxf5 {(0-0) 0.38/3 0} Qxf5 {-0.33/1 0} 7. a3 {(Qd3) 0.28/3 0
} Qg4 {(g6) -0.35/1 0} 8. O-O {0.26/3 0} e6 {0.09/1 0} 9. h3 {(Nc3) 0.32/3 0}
Qf5 {(Qg6) 0.15/1 0} 10. Re1 {(Nc3) 0.26/3 0} Be7 {(d5) -0.14/1 0} 11. Nc3 {
0.40/3 0} O-O {(Ne4) 0.06/1 0} 12. e4 {0.70/3 0} Qh5 {(Qg6) 0.17/1 0} 13. Bg5 {
(Qd3) 0.48/3 0} h6 {(d5) 0.05/1 0} 14. Bf4 {(Be3) 0.50/3 0} g5 {(e5) 0.15/1 0}
15. Bg3 {(Be3) 0.90/3 0} g4 {(Qg6) 0.29/1 0} 16. hxg4 {1.08/3 0} Nxg4 {
(Qxg4) 0.37/1 0} 17. d5 {1.14/3 0} Na5 {(exd5) 0.49/1 0} 18. Qd3 {
(Nh2) 1.50/3 0} b6 {(Bf6) 0.80/1 0} 19. dxe6 {(Nb5) 1.62/3 0} fxe6 {0.64/1 0}
20. b4 {(Nb5) 1.58/3 0} Nc6 {(Nb7) 1.21/1 0} 21. Qc4 {1.48/3 0} Nce5 {
(Nge5) 1.39/1 0} 22. Qxe6+ {1.56/3 0} Qf7 {1.39/1 0} 23. Qxf7+ {1.74/3 0} Kxf7
{(Nxf7) 1.45/1 0} 24. Nxe5+ {1.80/3 0} Nxe5 {1.33/1 0} 25. Bxe5 {1.86/3 0} dxe5
{1.36/1 0} 26. Nd5 {(Red1) 1.50/3 0} c6 {1.01/1 0} 27. Nxe7 {(Ne3) 1.04/3 0}
Kxe7 {0.90/1 0} 28. c4 {(f3) 0.96/3 0} Rad8 {(c5) 0.63/1 0} 29. Rad1 {1.06/3 0}
Rd4 {(b5) 0.52/1 0} 30. Rxd4 {(c5) 0.96/3 0} exd4 {0.47/1 0} 31. Rd1 {1.04/3 0}
c5 {0.47/1 0} 32. bxc5 {(f3) 1.00/3 0} bxc5 {0.42/1 0} 33. f3 {0.64/3 0} Rb8 {
(Ke6) 0.30/1 0} 34. Kf2 {0.66/3 0} Rb2+ {0.14/1 0} 35. Kg3 {0.51/3 0} Kf6 {
(Re2) 0.38/1 0} 36. Rh1 {0.52/3 0} Kg6 {(d3) 0.18/1 0} 37. f4 {(a4) 0.90/3 0}
Ra2 {(Rb3+) 0.37/1 0} 38. a4 {(e5) 0.42/3 0} Rxa4 {-0.27/1 0} 39. f5+ {
(Rc1) 0.00/3 0} Kg5 {-0.64/1 0} 40. Rc1 {-0.94/3 0} Ra3+ {(d3) -0.69/1 0} 41.
Kf2 {-0.75/3 0} Kf4 {-0.78/1 0} 42. f6 {(Re1) -0.88/3 0} Ra2+ {(Kxe4) -0.11/1 0
} 43. Kg1 {(Kf1) 0.36/3 0} Kg3 {-0.50/1 0} 44. e5 {(Kh1) 0.47/3 0} Rxg2+ {
(a5) -0.46/1 0} 45. Kf1 {0.06/3 0} Rf2+ {(d3) -0.60/1 0} 46. Kg1 {-0.22/3 0} d3
{(Re2) -0.60/1 0} 47. Rc3 {-0.45/3 0} Rf3 {(Rd2) -0.89/1 0} 48. f7 {
(Rc1) -0.61/3 0} Rxf7 {-0.64/1 0} 49. Rxd3+ {-0.87/3 0} Kf4 {-0.85/1 0} 50. Rd5
{(e6) -1.22/3 0} Re7 {(Rg7+) -0.86/1 0} 51. Rxc5 {-1.30/3 0} Rxe5 {-0.87/1 0}
52. Rc6 {-1.16/3 0} h5 {-0.79/1 0} 53. Ra6 {-1.22/3 0} Rc5 {(Re1+) -0.71/1 0}
54. Rxa7 {-0.62/3 0} Ke4 {-0.49/1 0} 55. Rd7 {(Rh7) -0.56/3 0} Ke5 {
(Rxc4) -0.49/1 0} 56. Kg2 {(Rh7) -0.38/3 0} h4 {(Rxc4) -0.29/1 0} 57. Rh7 {
(Re7+) -0.16/3 0} h3+ {(Rxc4) -0.09/1 0} 58. Kxh3 {(Kf3) 0.00/3 0} Rxc4 {
0.00/1 0} 59. Rh5+ {(Rf7) 0.00/3 0} Kd4 {Acepto las tablas 0.00/1 0} 1/2-1/2
and change colors... or sides.
rybka 3 human versus chesstiger 14
[Event "PC-820D-OLIV, 1Ply / 1Ply"]
[Site "PRIVADA"]
[Date "2008.09.05"]
[Round "1"]
[White "Rybka 3 Human 1-cpu 32-bit kz"]
[Black "Chess Tiger 14.0"]
[Result "1/2-1/2"]
[ECO "A06"]
[WhiteElo "2600"]
[BlackElo "2600"]
[PlyCount "114"]
[EventDate "2008.??.??"]
[TimeControl "120"]
1. Nf3 {Ambos ultima jugada del libro 0.09/1 0} d5 {-0.36/3 2} 2. d3 {
(Nc3) 0.18/1 0} e6 {(Nf6) -0.28/3 0} 3. Nc3 {0.28/1 0} Nf6 {(Nc6) -0.20/3 0} 4.
Be3 {(e3) 0.24/1 0} Nc6 {-0.42/3 0} 5. Nb5 {(d4) 0.09/1 0} a6 {-0.58/3 0} 6.
Nbd4 {0.09/1 0} Nxd4 {(Ng4) -0.54/3 0} 7. Bxd4 {-0.10/1 0} Bb4+ {
(Bd6) -0.40/3 0} 8. c3 {0.07/1 0} Bd6 {-0.20/3 0} 9. g3 {0.01/1 0} c5 {
(0-0) -0.24/3 0} 10. Bxf6 {(Be5) -0.10/1 0} Qxf6 {-0.42/3 0} 11. Qd2 {
(Bg2) -0.20/1 0} b6 {(Bd7) -0.86/3 0} 12. Bg2 {-0.06/1 0} h6 {(0-0) -1.06/3 0}
13. O-O {(a3) 0.00/1 0} Bb7 {(0-0) -1.10/3 0} 14. d4 {(Rfc1) -0.04/1 0} Bc6 {
(0-0) -1.04/3 0} 15. a4 {(Rac1) 0.02/1 0} O-O {-1.06/3 0} 16. a5 {(b3) 0.01/1 0
} bxa5 {(b5) -1.14/3 0} 17. Rxa5 {0.11/1 0} cxd4 {(Bb5) -0.60/3 0} 18. Nxd4 {
0.17/1 0} Bb7 {-0.38/3 0} 19. c4 {(Rfa1) 0.20/1 0} Be5 {(Bc7) -0.76/3 0} 20. e3
{0.29/1 0} Rfc8 {(Bc7) -0.56/3 0} 21. cxd5 {0.44/1 0} exd5 {-0.02/3 0} 22. Bxd5
{0.44/1 0} Bxd5 {(Rab8) 0.38/3 0} 23. Rxd5 {0.70/1 0} Rd8 {(Rab8) 0.46/3 0} 24.
Nc6 {(Qa5) 1.80/1 0} Re8 {(Rxd5) 1.12/3 0} 25. Nxe5 {1.16/1 0} Rxe5 {1.06/3 0}
26. Rc1 {(Rd6) 1.25/1 0} Rxd5 {(Rb8) 0.44/3 0} 27. Qxd5 {0.95/1 0} Rd8 {
0.70/3 0} 28. Qb7 {(Qa5) 0.90/1 0} Rd2 {0.70/3 0} 29. f4 {(Rc8+) 0.35/1 0} Rxb2
{-0.26/3 0} 30. Rc8+ {(Qc8+) 0.25/1 0} Kh7 {0.04/3 0} 31. Qe4+ {0.21/1 0} g6 {
(Qg6) -0.04/3 0} 32. Rc7 {(Rc6) 0.07/1 0} a5 {(Re2) -0.38/3 0} 33. Qd5 {
(Ra7) 0.10/1 0} Rb1+ {0.00/3 0} 34. Kg2 {0.27/1 0} Kg7 {(Rb2+) 0.00/3 0} 35.
Qxa5 {0.14/1 0} Rb2+ {0.12/3 0} 36. Kf3 {0.24/1 0} Rxh2 {0.26/3 0} 37. Qd5 {
(e4) 0.23/1 0} Rh5 {(h5) 0.46/3 0} 38. Qd4 {(Qd7) 0.31/1 0} Qxd4 {
(Rh2) 0.32/3 0} 39. exd4 {0.40/1 0} Rh3 {(Rb5) 0.64/3 0} 40. d5 {0.70/1 0} Kf8
{(Kf6) 1.04/3 0} 41. d6 {1.18/1 0} Ke8 {(f5) 1.00/3 0} 42. Kg4 {(Re7+) 1.24/1 0
} Rh2 {0.92/3 0} 43. Re7+ {1.01/1 0} Kf8 {0.82/3 0} 44. Rb7 {0.81/1 0} f5+ {
(Ke8) 0.31/3 0} 45. Kf3 {0.68/1 0} Rd2 {0.28/3 0} 46. Rb6 {0.28/1 0} Rd3+ {
0.16/3 0} 47. Kf2 {(Kg2) 0.02/1 0} Ke8 {-0.46/3 0} 48. d7+ {0.00/1 0} Kxd7 {
-0.40/3 0} 49. Rxg6 {0.01/1 0} Rd6 {0.28/3 0} 50. Rxd6+ {(Rg7+) 0.14/1 0} Kxd6
{0.20/3 0} 51. Ke3 {(Kf3) 0.15/1 0} Kc5 {(Ke6) 0.04/3 0} 52. Ke2 {0.00/1 0} Kd4
{(Kd5) -0.10/3 0} 53. Kf3 {0.04/1 0} h5 {(Kd5) -0.18/3 0} 54. Kf2 {
(Kg2) -0.57/1 0} Ke4 {-0.30/3 0} 55. Ke2 {(Kg2) -0.29/1 0} h4 {-0.34/3 0} 56.
gxh4 {0.00/1 0} Kxf4 {0.00/3 0} 57. h5 {0.00/1 0} Kg5 {
(Ke4) Acepto las tablas 0.00/3 0} 1/2-1/2
0.5-0.5
0.5-0.5
very surprising. this doesn't explain for that rybka is unable to win to ch14, 5 years old, and for that loses with ch15
postcript: if we try to evaluate this same experiment in depth 2 or 3, talvez is too complicated if we don't have information on the main algoritm, keeping in mind that many engines will be calculating ~5 ply, it is quite difficult to deduce something logical... also how many more plys we analyze more possibilities we will have of altering the experiment.
Don't made a mistake, it is not to obtain games of great quality, but of observing the differences of style and knowledge you position them and strategic of the engines.
The experiment has been useful and amusing.
a greeting.
1 - this evaluation is very useful to compare the evolution of different versions of the same engine, to compare among ryka 1, rybka 2, rybka 3 stiles, rybka mp. and to compare among fritz 6 to fritz 11, and versions deep... and to compare chesstiger versions... and to compare hiarsc 9 to 12...
2 - when we block analysis to 1 ply, at least we make sure of understanding the first movements candidates of the program, without influencing the quick cpus and code optimizations, perhaps be not exactly 1 ply, but it will always be the same analysis depth, and that is the important thing..
3 - this will reveal concept shortcomings in some positions...
4 - it can improve the level of an engine improving the selections that he/she makes in the first play... (first movements candidates) because starting from here the whole tree of variants is born.
5 - other engines calculate more +1 ply too...
bye. from spain. oliver.
Let go of the paranoia and simply address the ISSUES discussed.